TBC Mage Theorycrafting - Page 92

◊ Vontre · 04/28/07, 10:02 PM

The arcane blast part of the cycle is automatically determined by the encounter duration. Frostboltx3 is impossible for most people, you must have really mastered stopcasting to be able to squeeze that in. You've managed to cut your casting latency below .2 seconds which is monstrous.

That being the case, I'm not sure how you're doing a 2x frostbolt cycle as that would involve forcibly reducing your cast speed, which is always bad. 3x frostbolt is not included because I think most people can't do it. If you somehow can, it's obviously going to destroy any other casting method because you have no "wait time".

I may add an option for it and also an "override" option for the arcane blast part, but that somewhat defeats the purpose of the spreadsheet. More arcane blasts will ALWAYS lead to increased dps and decreased efficiency. The spreadsheet currently takes your encounter duration and automaticaly assigns an optimal number of arcane blasts based on how long you are chain-casting. Using that you can garner a good idea of how mana stats affect your dps, and also how much arcane blasting you can get away with.

Yeonora · 04/28/07, 10:53 PM

Yes, that's why ABx3 FBx2 comes in so clearly inferior. Even without using /stopcasting but "normal" FB- and AB-buttons, I always have to wait almost a whole second after the second frostbolt to start casting the first arcane blast of the next cycle.
Well, don't mind me asking for the ABx2-cycle, I finally figured out how to force that with encounter duration, sorry for being slow

But your answers for the other three questions would greatly interest me (think it's practical to just re-paste them, as it's a new page, at least for me).

1) How much dpm do I have with different spell usage / rotations? A dpm field would be nice.
4) When I switch "Mana Regen Consumables" to yes on the Main Sheet, it tells me that all my total dmg for fireball, scorch etc. goes up, but my total frostbolt damage stays exactly the same (the time goes up for alle spells). Also the arcane blast cycle stays the same in total damage. Why?
edit: yet another big one
5) Did you account for level-based non-penetrable resists? Based on posts like this one (although it's quite old), I assume lvl73 mobs have 24 innate resistance vs everything that is not binary. Which means that arcane hits, arcane crits and fire hits get shaved at least 5% of their damage (70 resistance is 15% mitigation at lvl 70, so 24 is 24/70*15% = 5.1429...%) in unmitigable level-based partial resists, while frost doesn't suffer from this problem due to frostbolt being binary and by some strange quirk in the game mechanics obviously not affected by those special resists. This also means that fire crits suffer twice, practically double-dipping on damage loss, because you only get an ignite from the non-resisted, non-absorbed damage, and then the ignite ticks can be partially resisted again, lowering the ignite to around 0.95*0.95=0.9025 of normal damage, meaning a fire crit gets 0.95*1.5*x+0.9025*0.6*x = 1.9665x, or 196.65% (instead of 210%) of normal hit damage for deep fire without spell power, or 0.95*1.75*x+0.9025*0.7*x = 2.29425x, or 229.425% (instead of 245%) of normal hit damage for arcane/fire with spell power. That's how I always did it in my spreadsheets at least, but none of them are as elaborated as yours of course, and I do not know for sure what exactly the level-based resists 70 vs 73 are.

Edit: You could add a [Frostboltx3] cycle and by default set it to off, like [Fireballx2], so people like me are happy

◊ Vontre · 04/29/07, 5:01 AM

Yeah I'll do the frostboltx3.

I reduced all damage by 6% to account for impenetrable level resist. This should affect frostbolt too by reducing chance to hit, I've never heard otherwise, magic resistance affects frost and fire equally.

Edit: Also ignite is no longer effected by damage multipliers on the target, so I'm not sure if it's damage dealt can be affected at all. I haven't noticed a resisted ignite since 2.0

Other question answers incoming. =)

◊ Vontre · 04/29/07, 5:20 AM

Mana regen consumables: the effect of these mp/5 consumables is very minor. The reason the arcane blast cycle damage is the same is because you end up using the same cycle. Frostbolt does the same damage because you do not go oom from chain-casting, so the extra mana does nothing. Fireball and other spells may see an increase in damage if the extra mana actually increases the number of spells you can cast.

So it's working fine =)

Yeonora · 04/29/07, 6:03 AM

Originally Posted by Vontre

I reduced all damage by 6% to account for impenetrable level resist. This should affect frostbolt too by reducing chance to hit, I've never heard otherwise, magic resistance affects frost and fire equally.

Edit: Also ignite is no longer effected by damage multipliers on the target, so I'm not sure if it's damage dealt can be affected at all. I haven't noticed a resisted ignite since 2.0

I admittedly have no idea if ignite still can be partially resisted by mobs without "normal" resistance but higher in level than you because I didn't have an ignite spec since 2.0. I'll try it out or ask the fire mage in my raid / browse through his swstats details.
But what I do know is that frostbolts certainly do not get 5% more resists than they should, meaning if I'm up at at least 16% hit (which is rather easy with an elemental shaman) fighting for example Gruul, my frostbolts show up in swstats as ~1% resisted, and of course no partial resists.
I will save the data next time and provide a snapshot (gotta pester the raid leaders so I'll get the elemental shaman again *g*).

Edit: This post is not the newest and is related to melee attacks rather than spells, but it certainly supports a 1-roll-system (meaning 5% crit always is 5% of your attacks crit, independent of hit rating).
Just stumbled upon this looking for proof and exact data on the non-mitigable level-based resist rates. I've been reading a lot about this and also that binary spells aren't affected by it, but just now I can't seem to be able to find it any more :-/

Another edit: Again, the best thread I can find about that topic is this one, although it's not one I remember reading.
Some important quotes:

Using only mobs for which I have now >1000 samples I'm quite confident the actual values are as follows:

Level 63: 0.94 = 24 resist
Level 62: 0.96 = 16 resist
Level 61: 0.98 = 8 resist
Level 60: 1.00 = 0 resist

Now if that stays the same at lvl70, we should be seeing less reduction from these level-based resists, because 24 at lvl 60 is 24/60*15% = 6% reduction, 24 at lvl 70 is 24/70*15% =~ 5.14% reduction.

An update on the frost data. I have now 768 samples on Anubisath Defenders, hit rate with +5% to hit is 0.99218, pretty close to the expected value for mob with 0 frost resistance. If it had 15 frost resistance as seems to be the case for base partial resists, then the hit rate would be 0.99 * 0.9625 = 0.9528. For this number of cases this seems to be a good indicator that binary spells indeed do not have this problem.

Found an easily kited level 61 named High Executor Derrington in horde AD camp in Tirisfal Glades. Used lvl 1 frostbolts. Binary spell gives that all spells are either resisted or give full damage.

Level 61
High Executor Derrington
2% to hit

630 frostbolts
20 resists
gives 3.15% of the frostbolts resisted

The miss chance against a level 61 is 5% and with 2% hit this gives a 3% miss chance that doesn't involve resistance.
If the level 61 would have 8 level based resistance to frost and the damage reduction would be the same as for non binary spells we should see 5% resists.

This is also one of the reasons I stuck with frostbolts instead of fire spells for raiding. They come out better than they appear on paper, not taking those peculiar resists into account.

Another edit:
I'm 100% positive that this quote from the thread is not true:

Binary spells (eg frostbolt)

For binary spells, the chance to hit is multiplied by the chance to land despite resistance, giving an overall chance to affect the target. Because this calculation is combined, there are only white resists against binary spells. The formula is as follows:

Chance = (base chance + talents + gear) * (100% - (max (0, min (300, base resistance + 5 x (target lvl - attacker lvl))) / 4))

This overall chance is capped at 99%. Note that this means that +% to hit gear will act like spell penetration gear for binary spells, but as stated above, not for non-binary spells. However, spell penetration gear will offer no benefit against a target with no resistance.

I will try to prove that spell hit does not counteract level-based resistances for binary spells, but that level-based resistances for binary spells are simply not taken into the calculation by trying to reach exactly 16% spellhit on our next Magtheridon- or Gruul-attempt, i.e. 1% from elemental precision, 1% from the draenei-aura, 3% from totem of wrath and as close as possible to exactly 11% spellhit on gear.

Kavan · 04/29/07, 7:57 AM

Originally Posted by Yeonora

Now if that stays the same at lvl70, we should be seeing less reduction from these level-based resists, because 24 at lvl 60 is 24/60*15% = 6% reduction, 24 at lvl 70 is 24/70*15% =~ 5.14% reduction.

My parsing for level 73 targets indicates that the reduction is a flat 2% per level difference, so not 8 resist per level, but I'll need some more data before I'm really convinced.

Yeonora · 04/29/07, 8:01 AM

This gets more interesting the more I try to nail it down.
Here is a post of the same player, but on the official forums and containing lots of different, additional information.

I will use a custom mod written for this purpose by this player (Kavan), to determine the resists of Gruul / Magtheridon: Resist Tracker
Also I'll use SW Stats which also shows partial resists, amount of partial resists, spell hits, complete spell resists and so on.
Hope this is sufficient.

Edit: Oh, hi Kavan, there you are

Did you collect some data to resolve this problem posted by you?

The reason why I believe this is not simply a feature but a bug is because binary spells are not affected by it. When no penetration is used the level of resists corresponds to the product of level-based miss rate and here described level-based resistances. However when penetration is used it is possible to counter the resistance part and achieve hit rate expected from miss rate alone. (it might be the case that binary spells don't get level-based resistances applied in the first place, I lack data to confirm this)

Meaning specifically: Do you have evidence if you need spell penetration for binary spells to reach 1% resist on bosses or is the level-based resistance not included into the calculation at all?

Nother edit: Nevermind the Blizzard post regarding the 1-roll-system I linked earlier. I read up on the spell theorycraft that indicates a 2-roll-system for spells here. Although no conclusive tests on same or +3 lvl mobs were made, assuming a 2-roll-system rather than 1-roll seems to be the wise choice atm.

Yeonora · 04/29/07, 2:06 PM

Some preliminary results, mainly to get some feedback on how to improve recording the data so it's most valid / informational.
This data is from the SW Stats segments for four Netherspite, one Prince Malchezaar and two Nightbane fights.
I managed to tweak my equipment so I got exaclty 139 hit rating, which equals 11.02% chance to hit in the character sheet, +1% EP, +1% Draenei Aura, +3% Totem of Wrath = 16.02%, which should result in getting 1% complete resists.

As you can see, 349 frostbolts with 3 resists.

Which results in 0.9% resists, as expected. The 538 partially resisted damage is from Chill of Norganna only (I had an eye on it and noticed the resists there while fighting Prince), that's the proc from Wrath of the Titans.

316 arcane blasts, 4 full resists.

Resulting in 1.2% full resists, as expected. Partial resists are at 2.8%, which is much lower than expected. Guess I need a much bigger sample size.

The last picture doesn't prove anything because of the low sample size, but it just looks so nice arcane getting attributed 24 resists and frost none vs. Prince Malchezaar

Kavan · 04/29/07, 4:14 PM

Hi Yeonora

You're probably the first in a long while that's been interested in my resist tracking mod, it's a bit of an obscure topic for most. Which means that the version available for download is a bit out of date, specifically it's lacking some of the buffs from shaman totems, so if you're frequently grouped with shamans be careful how you interpret the data (it'll be recorded under the lower +to hit value). I'll try to get to it later today and upload the latest version.

Regarding the question about binary spells I don't have anything to contribute. I'm a pure arcane mage, so it's all partial resists for me. The last time I used frostbolts was in AQ40 :P

Yeonora · 04/29/07, 6:03 PM

That would explain the negative numbers for frost resistance across the board. Well, we know that it's supposed to be 0 resistance *g*
Some more info I collected today on Hydross (we'll do Gruul + Mag on Wednesday):

No frost resists, 28 arcane resist with 128 samples calculated, 30 fire resist with 109 samples calculated.
I was using an AB/Fireball cycle for the frost phase and an AB/Frostbolt cycle for nature.

At least we can see that frost indeed does not suffer from level-based resists, while arcane and fire does. I was again using the perfectly matched equipment for 11.02% hit on equip +4% shaman +1% EP, so we cannot attribute the "missing" frost resists to me having "too much" hit and thus somehow overriding the level-based resists as was claimed.

Kavan, do you think the calculated resist numbers for arcane and fire are too low due to Resist Tracker not accounting for the extra +hit from the Shaman, or does it just push frost resists into the negative?

chase · 04/29/07, 7:08 PM

Do you think you could post a link to your combat log for a boss fight? Possibly several?

The results of the mods seem a bit off.

Kavan · 04/29/07, 10:46 PM

It'll just shift the displayed data, because it tries to intepret the raw resist information given known miss rate collected at recording time. You can get to the raw data by using the /rt resistdetail command.

The most data I have collected is from Prince Malchezaar, this is with 30 penetration and 65 hit rating + 10% from arcane. This would give you theoretical 1.85% miss rate. Based on 2338 samples I have:

full resist: 1.71%
75% resist: 1.54%
50% resist: 5.47%
25% resist: 10.69%
no resist: 80.58%

Partial resists account for a factor of 0.9332 (value from 1780 samples on Shade of Aran give a factor of 0.9384). As I said, not really enough data to distinguish 2% per level vs 8 resist per level, but so far the data is in favor of 2%.

EDIT: I uploaded new version to curse that supports Totem of Wrath.

Yeonora · 04/30/07, 9:06 AM

Originally Posted by chase

Do you think you could post a link to your combat log for a boss fight? Possibly several?

The results of the mods seem a bit off.

I what way? And do you relate to Resist Tracker or SW Stats?
I'll make sure to figure out how to save my combat log for the next Gruul/Mag fight just in case

And thanks Kavan for the new upload!

Groglox · 04/30/07, 3:51 PM

This might be slightly offtopic, but I figure mages would know the answer better than general UI people. I picked up nightstaff of the everliving last night for evocation, and I set my event in itemrack to use my "evoc" set which has my "of spirit" wand and the staff, and it does work. One thing I noticed though is when it goes to requip my normal damage set afterwards, it doesnt reequip my offhand.

Has anyone else encountered this or have a solution?

Tempestra · 04/30/07, 4:25 PM

While I can't give you a soultion regarding setting up the event in itemrack, I have had similar problems in the past.

I use closetgnome and use hotkeys to swap before and after an evoc - haven't had any problems whatsoever. It's a bit more manual, but I'm not really doing anything else during the evocation anyways =)

Cryic · 04/30/07, 5:29 PM

Originally Posted by Groglox

This might be slightly offtopic, but I figure mages would know the answer better than general UI people. I picked up nightstaff of the everliving last night for evocation, and I set my event in itemrack to use my "evoc" set which has my "of spirit" wand and the staff, and it does work. One thing I noticed though is when it goes to requip my normal damage set afterwards, it doesnt reequip my offhand.

Has anyone else encountered this or have a solution?

I've ran into this a few times. I believe it's a somewhat rare occurrence when you are switching from a 1h+offhand -> Staff for evoc with Itemrack. After the evoc has ran it's course, it can sometimes try and equip your offhand a split second before the 1hander resulting in an error (the 1hander has to be equipped first).

I've also ran into problems with itemrack going completely haywire if my bags are full and I evoc. Even after everything is done and bag space is freed up.

Anyways, to solve this little problem, I normally tap my itemrack hotkey for my dps set after evoc is finished. Most times nothing happens, but once in a while i'll get a pause on my spellcasting where the item is being equipped.

Cloudgatherer · 04/30/07, 5:37 PM

Originally Posted by Groglox

One thing I noticed though is when it goes to requip my normal damage set afterwards, it doesnt reequip my offhand.

What version of itemrack and does your re-equip set specify your offhand item? I've been doing exactly the same thing, sans a wand, and have had no problems. I'll try throwing in a wand this evening and seeing if I get the same problem.

chase · 04/30/07, 6:17 PM

Originally Posted by Yeonora

I what way? And do you relate to Resist Tracker or SW Stats?
I'll make sure to figure out how to save my combat log for the next Gruul/Mag fight just in case

And thanks Kavan for the new upload!

Negative resists don't lead to a large confidence in the tracking done by the mods.

To start logging your combat log you simply type /combatlog before and after the fight. The resulting file will be in your wow folder in a folder called Logs. (Must /console reloadui or logout to get the file to write)

Yeonora · 04/30/07, 6:30 PM

Originally Posted by chase

Negative resists don't lead to a large confidence in the tracking done by the mods.

To start logging your combat log you simply type /combatlog before and after the fight. The resulting file will be in your wow folder in a folder called Logs. (Must /console reloadui or logout to get the file to write)

Well, we already figured out that's because Kavan's old version of Resist Tracker didn't recognize and take into account the additional %hit provided by the Draenei Shaman's aura and his totem of wrath.
I already got the updated version and will use that henceforth... plus the SW stats statistics are also there and certainly reliable.

Kavan · 04/30/07, 8:31 PM

Originally Posted by chase

Negative resists don't lead to a large confidence in the tracking done by the mods.

You have to understand what the mod is displaying. Internally the mod is tracking resists for each mob and keeping separate record for each variant of hit rating/extra hit% from talents and buffs/penetration/mob name/mob level. To display the resists it uses two methods, one is ignoring full resists (assuming they are all from misses), the other uses the theoretical miss rate to adjust the data (more accurate for binary spells and high resist mobs, but needs more data to get the same significance levels). This means that if the ammount of misses is higher than expected based on level difference and hit rating, the resist displayed is negative. Now everyone can display a 0 in this case, but to me this value is meaningful as it tells me how much variation to expect at a certain number of samples.

Zeku · 05/01/07, 2:43 AM

It's just as easy to make a macro if there's problems.

/equip [button:1] staff of spirity stuff;
/equip [button:1] wand of spirity stuff;
/equip [button:2] weapon of kicking butt;
/equip [button:2] offhand of kicking butt;
/equip [button:2] wand of kicking butt;
/cast [button:1] evocation;

You don't have to use the entire name of the item, just get a few of the first characters right. Left click to evo, right click to finish.

It will fail in a battleground for some reason, but you just have to mash it a bit and it still works.

◊ Vontre · 05/01/07, 3:39 AM

Originally Posted by Zeku

It's just as easy to make a macro if there's problems.

/equip [button:1] staff of spirity stuff;
/equip [button:1] wand of spirity stuff;
/equip [button:2] weapon of kicking butt;
/equip [button:2] offhand of kicking butt;
/equip [button:2] wand of kicking butt;
/cast [button:1] evocation;

You don't have to use the entire name of the item, just get a few of the first characters right. Left click to evo, right click to finish.

It will fail in a battleground for some reason, but you just have to mash it a bit and it still works.

I want actual items named that.

Zeku · 05/01/07, 7:11 AM

On the professions forum, there was (one, as far as I know) post that tailor sets were nerfed on PTR. Somewhere around -37 damage, +24 int for the full spellfire set. This has since exploded across nearly every forum. I don't know anyone who has personally confirmed it yet, it would be funny if it was a prank.

That would hurt fire slightly more than arcane. I would feel sorry for the people who have already done all the farming and crafting.

Anduryondon · 05/01/07, 7:14 AM

worldofraids.com are the actual stats, nearly every piece +8int and -11 - -12 spelldmg

Antiphonal · 05/01/07, 9:37 AM

Confirmed.

I think it hurts spellfire tailors a bit more since two of the three items now have immediate sidegrades - Girdle of Ruination (actually an upgrade) and Handwraps of Flowing Thought. Because mages scale so well with crit and hit and can wear and benefit from a large variety of pieces, the Spellfire changes make them roughly Tier4 in quality.

The bad part is what happens to the affliction warlocks (especially Drain-spec affliction) and shadow priests. They don't need crit. They need minimal +hit. They don't have a mechanic to scale with +int. The Frozen Shadoweave set was THEIR set. It had everything they wanted, nothing they didn't. They get upgrades in Tier6, but they have to endure gear with tons of wasted/useless crap on it (spirit, crit, etc). So basically the firelock or destro-shadowbolt warlock (shadowmage) gets quite a bit more benefit out of the tier gear than the classic DoT and Drain setup.

Shadow Priests don't even have the option of speccing into a tree that scales well with +crit. Now that SW-D has twice as long a cooldown, their spell rotation will have even fewer spells that can crit. Ick.

All that said - the situation was starting to remind me of back when I farmed the pattern and mats for the Robe of the Archmage. It was clearly and obviously better than everything else all the way to the Tier2.5 robe from C'Thun. That tailoring epic would take you all the way from LBRS to C'Thun - and mayber further since the mana clicky became a play-defining little trick. It, like Spellfire and FSW, was itemized PERFECTLY. Had everything that was wanted, omitted everything superfluous.

I've had my set for a couple months now, and I don't think I will feel cheated because I can see how necessary this is. FSW will still fulfill its function. You still can hit 70 as a frost mage, affliction warlock, or shadow priest, and have 6/8 epic tailored gear (as well as a great tailored cape) and compete on the damage meters with the long-time raiders. It just won't be a clear case of "Go tailoring or suck" which is what we have now.